The use of AI in fraud detection has been picking up pace in the past few years due to its ability to detect patterns in data that surpass human capabilities. And with the rise of Machine Learning (ML) and Deep Learning solutions, these algorithms are becoming more advanced.
In their report on the challenges and opportunities of new technologies for Anti-Money Laundering (AML) and Combating the Financing of Terrorism (CFT), the Financial Action Task Force (FATF) acknowledged several benefits of these new technologies, including:
Enabling faster and more accurate data collection, processing, and analysis to help supervisors identify, assess, monitor and communicate suspicious transactions more effectively and in near real time
Facilitating communication between different actors in the financial ecosystem
Improving identification, understanding, and management of money laundering and terrorist financing risks
But they also raised some concerns related to operational and regulatory constraints and complexities associated with updating legacy systems, explainability and interpretability.
According to FTAF, smaller financial institutions, in particular, often lack internal capacity or confidence to evaluate the effectiveness of a given innovative solution among a large and growing range of competing vendors and products. It’s challenging to determine whether the solution is appropriate for the institution’s risk profile, customer base, and business activities, but also how to implement ML models and manage the associated risks.
What should the first step be?
Start with a Thorough Analysis
Before starting any project on adopting new technologies for AML fraud detection, we advise financial institutions to conduct an analysis covering five levels:
Understand the unique business model of the company and the associated risks: the specific vulnerabilities based on the current operations, products, services
Evaluate existing compliance processes and governance structures: AML policies, precedures and oversight mechanisms
Audit the current IT infrastructure and technological setup
Assess data readiness: the quality, quantity, accessibility, structure of the data, and data protection measures
Consider the HR aspect: the team’s current skillset, training needs, resistance to change
10 Key Questions to Ask
Based on our experience implementing AI & ML solutions for AML fraud detection, here are ten questions to help you kick off this analysis:
What are the strengths and weaknesses of the current Financial Crime Compliance (FCC) compliance process?
Did specific risk areas increase in the last couple of months or years?
What is the proportion of false positives?
How much time does the investigation team spend evaluating false positives compared to high risk transactions?
Does a risk-based approach segregate investigation effort from potential ratings? And can you implement the risk-based approach homogeneously in the company?
Is your rules engine flexible enough to cover the latest regulations? What about tracking new changes in customer behavior?
Are you evaluating data quality and availability? Which effort is required to collect and clean transactions and prospect and client data?
What is the ratio of manual data checks vs automated checks?
What is the team's readiness to adopt new technologies?
How do you intend to introduce new technologies? By gradually adding new components? Or are you considering replacing parts of your legacy systems?
Identifying weaknesses and strengths at these five levels will lead to a list of requirements for designing an effective solution. It could involve data enhancement, practice redesign or upgrading existing IT systems with the latest technologies.
Concerns related to the cost of new technologies, the ability to ramp up resources for the implementation of these new technologies, as well as the cultural shift are also aspects to factor in while assessing the feasibility of a new solution.
Towards Augmented Compliance
Like all emerging technologies, it's essential to weigh the costs and benefits of AI before integrating it into your compliance processes. We must remember that AI isn't an isolated technology; its true value lies in enhancing efficiency, accelerating processes, reducing costs, and freeing-up resources. If it doesn't contribute to a more robust compliance framework, then we're missing its true potential.
Subscribe to our newsletter
You’ll receive insights, strategies, and best practices that help you succeed in adopting and implementing AI & Data. Only what matters. Once a month.